Safety planning for live game operators, beyond the fire exit
Passing an inspection means your building met a checklist on a particular morning. Being prepared means that when something goes wrong at half past nine on a Saturday, with a full house and your newest hire on shift, the right thing happens without anyone having to work out what the right thing is.
Those are different projects. This one is about the second. Nothing here is legal or insurance advice, and your own advisers and local requirements come first; what follows is the operational side, the part that lives in habits rather than in a certificate.
The incidents that actually happen are mundane
Ask operators who have been running for years and the stories are remarkably consistent. Someone has a panic response in a dark room. Someone faints, or has a medical episode that has nothing to do with your game. Someone catches a shin on a set piece in low light, or trips on a cable. A mechanism jams and a group is behind a door that will not open. A staff member is alone in the building at closing with a group who have been drinking.
None of that is dramatic. All of it is foreseeable, which means all of it has a preventable version and a practised response.
Egress that works when a player panics
The design question is not whether there is an exit. It is whether a frightened person, in the dark, who has never been in your building before, can get out without help and without solving anything.
- Every room needs a way out that requires no puzzle, no code and no staff. Push to open, obvious, and in the direction someone would instinctively move.
- The exit route must be visible in the dark from wherever players actually are, including from the floor.
- Say in the briefing, in plain words, where the exit is and that using it ends the game with no penalty and no fuss. People do not use an exit they think will embarrass them.
- Walk each room after any set change with the lights at game level and check the path is still clear. Set dressing migrates.
Every locked mechanism needs a fail-safe, and a schedule
Anything that can hold a person or hold progress needs a manual override that works with no power and no network, and every member of staff needs to know where it is and have used it. Magnetic locks that release when power drops, mechanical key overrides on any electronically controlled door, and a physical master key set kept somewhere staff can reach quickly rather than somewhere secure enough to be useless.
Then test on a schedule, written down, with a date and an initial against each item, because untested fail-safes are a belief rather than a control. The weekly pass covers the things that fail quietly: exit routes clear, emergency lighting working, first aid kit stocked, cameras and audio live in every room, radios or intercoms charged. The monthly pass is the one people skip: actually operating every manual override, actually triggering the emergency release, actually checking the battery in the prop that has never once failed.
Watching, briefing and logging
The briefing lines that prevent injuries
A good safety briefing is short, specific and framed around the game rather than around liability. Long disclaimers are ignored; four concrete sentences are not.
- Nothing in this room requires force. If it does not open, it is not the answer.
- Do not climb, stand on furniture, or move anything marked as fixed.
- Here is the exit, here is how you call us, and you can stop at any time.
- The room gets dark or loud at this point, so you know in advance.
The first line prevents more damage than any other single thing you will say, to both your props and your players. The last one prevents most panic responses, because what frightens people is being surprised rather than being in the dark.
Notice distress in seconds
Monitoring is a safety control, not just a hint-delivery system. Whoever is watching should be able to see and hear every group continuously, with no blind corner, and should be watching rather than doing paperwork in another room. Cameras that can see in the actual game lighting, audio that is on rather than muted, and a way for a player to call for help that does not require finding a button they were told about forty minutes ago.
Teach staff the signals: someone standing apart from the group, someone sitting down, breathing that changes, a person asking twice to leave in a joking tone. The response is the same every time, and it should be a rehearsed line rather than an improvisation: lights up, door open, calm voice, no clock, no questions asked in front of the group.
Write down the near misses
Keep an incident log and make it a normal, blame-free habit rather than a thing that only appears after something bad. Log the actual incidents, obviously. But the real value is in the near misses: the lock that stuck and then opened, the group who could not find the exit in the debrief walkthrough, the prop that came off the wall between sessions.
Every serious incident in any operation has a queue of near misses in front of it. A log is how you see the queue. Review it monthly with the same seriousness as the takings, and let anyone on the team add to it without asking permission.
The end of the night, and staying competent
Lone working and closing
The end of the night is when staffing is thinnest and risk is highest. Decide your minimum staffing and hold it, including for the last session. Agree a check-out routine so someone knows a shift ended safely, whether that is a message to a manager or a call. Make sure a lone member of staff can summon help without leaving the desk, and that they know they may end a session, refuse entry, or lock up early if a group is intoxicated or aggressive, without needing to ring you first for permission.
Put that authority in writing. Staff hesitate when they are not sure they are allowed to.
Practise, briefly, often
Competence decays. A quarterly ten-minute drill, run at the start of a shift, keeps it alive: release a group with the manual override, walk the evacuation route, locate the first aid kit and the emergency contacts, run the script for a panicking player out loud. Every new hire does all of it before their first solo shift, and every existing member of staff does it again this quarter, including the ones who have been there for years and especially the ones who are confident.
What changes outdoors
If any part of your operation happens outside the building, the risk list is different rather than smaller. Traffic and road crossings. Weather, heat and cold. Uneven ground and poor light at the times people play. Members of the public who are not in your game. Above all, players you cannot see, which removes the monitoring control you rely on indoors and replaces it with design decisions made before anyone plays.
That is the discipline we apply when City Sleuth partners build a route: stops chosen to be safe at the hours tickets are sold, no puzzle that rewards stepping into a road or climbing anything, clear guidance in the app about crossings and terrain, and a way to reach a human when something goes wrong. The control moves from a game master watching a screen to choices baked into the route, which means the route has to be walked at the worst hour you would sell rather than at noon on a pleasant afternoon.
What to do this week
Pick one room, turn the lights down to game level, and time how long it takes someone who has never played it to find and use the exit. Then open your manual override, in front of whoever is on shift, and see whether they knew where it was.
Whatever those two tests tell you is your actual safety position, and it is more informative than any certificate on the wall. If part of your plan involves taking players outside the building, the City Sleuth partner overview describes how an unstaffed route is designed around exactly these questions.